Links to mobile versions on the 1st page.

 

Fuck you! No Candy Crush graphics! (At least until the open-source version, then texture packs should be viable).

Release date is tomorrow (18th July) for Android and iOS.

Ending will also be at Rezzed in Birmingham this weekend as a part of the Leftfield collection along with Icefishing 5, Trash TV, etc. Come check us out.

Note to self: update desktop versions to check stage.displayState

Release date moved to 18th June for mobile releases. Still struggling with Android, but getting closer to narrowing down what is causing the issue (empty AIR projects are accepted by the store but Ending isn't - so it's trial and error time...)

They're just rectangles, no rotation ever.

Technically I must visit every rect/body once. They are all of uniform size (width and height different). They can overlap, and often compress into one unit's worth of space (allowing you to hold up and quickly create little ant-towers of dudes climbing up each other).

[because they are often all in the same bucket - so multiple buckets is pointless]

Spatial hashing won't work because the swarm entities can overlap - they end up in the same bucket nearly all the time. I wouldn't gain any performance increase.

Quadtrees only make sense for items that vary in size - otherwise maintaining the quads simply eats up all the cpu you were trying to save. I know this because I've implemented quadtrees before - they're fucking useless for almost everything.

I've written physics for hashing and quadtrees, trust me when I say I know what I'm talking about and that it won't offer any performance increase in this instance.

What I need is counter-intuitive to a hash-bucket system because they are often all in the same bucket - so multiple buckets is pointless.

I'm making a game where you control a swarm. A bit like Enough Plumbers meets Pikmin.

They can stand on each other's heads and climb up each other though, which creates some solutions to obstacles that's nicely like the wave of zombies in that War Z trailer that was doing the rounds.

Of course I have to verify that a swarm member is overlapping another swarm member. At the moment I'm doing it the brute force way (just prototyping at the moment). I'm wondering what I could do to optimise the operation.

I thought of doing a spatial hash but they tend to clump in one location a lot - so I won't see the benefit of a system that can split the checks into groups.

Noooo, the image on the front page is gone. It was a dancing gif of Hitler timed perfectly with the song.

 

http://forums.tigsource.com/index.php?topic=30169.0

There's a thread like this on another forum I'm on.

It's amazing: http://yayhooray.net/thread/6991/

Built the Flash on Windows.

Just spent the past few hours building it on every OS for every OS.

Which reminds me, I've still the Android store to upload to. 

Desktop downloads for the big 3 operating systems now available on the first page, along with the other release targets planned.

Without mid-air control you don't get to create "wiggle-jump" tasks for the player to perform.

That's not to say that restrictions aren't good for a game. They just make it harder to design levels for.

I'll be releasing standalones with the App Store release.

I just got notice that it's been approved, should go live on 11th June. Hopefully I should be able to get the Android version sorted by then as well.

grats dude 

Submitted to the app store. Now waiting for the results of the review. I'm going to try out an Android build now.

The version at the link is pretty much the final game. Consider this done. I'll be updating with platform releases and bug fixes from now on.

note to self: mouse-pointer isn't hiding when fullscreen

More UI work 

Going to look at this scaling crap on iOS now.

To be honest, no.

The source code needs to be tidied up for porting to other languages anyway. That means converting to using a sprite sheet - which will mean people can design their own graphics.

Added sound toggle and fullscreen toggle to the browser version (plus cleaned up the page).

More balancing on Adventure mode - which I now dub Gauntlet mode, because it's now 9 rooms minimum of pain.

One more round of live-testing before I submit the iOS version.

note to self: save state isn't recording the current score